Large-scale biological datasets reveal increasingly complex genetic and protein-protein interaction networks. As a consequence of this complexity, for key proteins with many interaction partners that are found at central positions in the network, it is difficult to extract quantitative information on how each interaction contributes to distinctor overlapping cellular functions, and, importantly, how changes to individual interactions result in altered function or disease. This knowledge would be critical for progress in many fields, such as biological engineering that requires predictive control of signaling networks, or the development of new targeted interventions in precision medicine. The long-term objective of this project is to advance studies that dissect complex protein networks by creating and testing a new multidisciplinary bioengineering approach that links systematic computational prediction of molecular perturbations at the amino acid-level to their effects on biological processes at the systems-level. The project will initially target the highly-conserved multi-functional Gsp1/Ran GTPase that controls key eukaryotic processes. The approach first engineers defined perturbations to protein-protein interactions by amino acid mutations (edge perturbations). The second step determines the functional effects of these perturbations at the cellular and organism level. The project advances technologies developed in three groups and integrates them into a platform that combines physics-based modeling and reengineering of interactions (Kortemme), mechanistic analysis of sequence-structure-function-fitness relationships (Bolon), and large-scale physical and genetic interaction mapping (Krogan). Innovative aspects are (i) the new integration of approaches and (ii) preliminary data indicating that the approach can not only dissect existing Gsp1 functions but also discover new biological functions, even for this well-studied protein. Aim 1 proposes to develop, test, and advance a validated computational model that can be used both to engineer and to interpret quantitative perturbations to interactions in protein-protein networks. Aim 2 will test hypotheses from Aim 1 by determining the consequences of engineered perturbations on cellular function in the model organism S. cerevisiae, chosen for its genetic tractability and extensive available information to validate the approach. Integration of the model from Aim 1 and data from Aim 2 will lead to an improved model and new hypotheses that will in turn be tested, resulting in new knowledge of the mechanistic basis of systems-level changes in function. Aim 3 will translate our platform into mammalian cells, which will provide fundamental insights into conserved and divergent mechanisms of Gsp1/Ran that is 83% identical in amino acid sequence between yeast and human. Our study will result in a validated technological platform that can be applied to other systems to derive predictive models of consequences of mutations on cellular function and organismal fitness. Long-term, we expect this platform to impact bioengineering approaches as well as the development of new targeted therapies that make precise network perturbations.